Figure 1.
The definition of Regions of Inerest (ROIs).
The brain is segmented into six lobes according to anatomy, including left frontal lobe, right frontal lobe, left temporal lobe, right temporal lobe, left occipital lobe and right occipital lobe. The six lobes are then classified into normal lobes and lesion lobe according to corresponding CT images for subsequent impedance observation.
Figure 2.
The procedure of EIT validation experiment in realistic head phantom.
Top Left: the six pre-defined positions in the phantom. Top Right: a photo of the system calibration experiment. Bottom: The EIT images of an agar cylinder at the six pre-defined positions in the phantom. Note that each image is displayed with its own colorbar in order to show the details.
Figure 3.
The result of EIT validation experiment in realistic head phantom.
The right shows spatial resolutions as a function of spatial location within the phantom. Position E and F (the two positions at 3 and 9 o'clock close to boundary) shows the best spatial resolution among the six positions. The left shows images intensity of the agar cylinders versus their known impedance at the six pre-defined positions. It can be seen that the intensity of EIT images for the agar cylinder is linearly related with its known impedance. The different slope for each spatial position indicates spatially-dependent system sensitivity.
Figure 4.
The normalized mean impedance before and after the mannitol injection over time.
When the normalized impedance change 30 min before the mannitol injection was chosen as a control dataset, the change 5 min after the injection as an experimental dataset, the impedance of brain significantly increases. The error bar represents SD.
Figure 5.
3 out of the 14 patients where, after an initial rise in impedance the impedance remained high.
Figure 6.
3 of the 9 patients where, after an initial rise in impedance there was a subsequent decrease in impedance.
Figure 7.
Comparison of impedance variation over time in ROIs with normal tissue versus those with lesion tissue in all 23 patients.
Left: the first patient group shows that normal tissue promptly increases with the mannitol injection and remains high impedance value after injection (14 cases). Right: the second patient group shows the similar result during the injection but the impedance of normal tissue decreases after injection (9 cases). Lesion tissue in both patient groups appears insensitive to dehydration treatment. The error bar represents SD.
Figure 8.
Comparison of EIT measurements and ICP on 5 patients before and after mannital injection.
The white bar represents the change of relative impedance, and the grey bar represents the change of ICP, before and after mannitol injection. The difference significance is set to P<0.01. The error bar represents SD.